Semantic Data Control

 —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  — —











Sup­port trans­la­tion: http://amzn.to/1Z7d5oc
 — —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  — —

Semantic Data Con­trol

Semantic data con­trol typ­ic­ally includes view man­age­ment, secur­ity con­trol and semantic integ­rity con­trol. These func­tions ensure that author­ised users per­form cor­rect oper­a­tions on the data­base. This con­trib­utes to data­base integ­rity.

Views in Dis­trib­uted DBMS

A view in a dis­trib­uted data­base is derived from frag­men­ted rela­tions stored at dif­fer­ent sites. When a view is defined, its name and its retriev­al query are stored in a cata­logue.

The map­ping of a query expressed on views is done in the same way as in cent­ral­ised sys­tems. In cent­ral­ized sys­tems it is done through query modi­fic­a­tion. A dis­trib­uted query is pro­cessed by the dis­trib­uted query pro­cessor.

Views in a dis­trib­uted DBMS can be done by using rules sim­il­ar to frag­ment defin­i­tion rules. The query pro­cessor maps the dis­trib­uted query into a query on phys­ic­al frag­ments.

Views derived from the dis­trib­uted data­base may be costly to eval­u­ate. View deriv­a­tion is done by mer­ging the view qual­i­fic­a­tion with the query qual­i­fic­a­tion. An altern­at­ive is to use snap­shots. A snap­shot rep­res­ents a par­tic­u­lar state of a data­base. This means that it is stat­ic. Snap­shots are use­ful when users are not inter­ested in see­ing the most recent ver­sion of the data­base. Access through snap­shots is adequate for quer­ies that are pre­defined.

Dis­trib­uted Author­iz­a­tion Con­trol

More com­plex than in cent­ral data­bases due to the fact that objects and users can be dis­trib­uted. These prob­lems are:
 — Remote user authen­tic­a­tion
 — Man­age­ment of dis­trib­uted author­iz­a­tion rules
 — Hand­ling of views and user groups.

To pre­vent remote user access by unau­thor­ized users, 2 approaches are used:
 — Inform­a­tion of author­ized users are stored on all sites in the cata­logue (more costly)
 — All sites of dis­trib­uted DBMS authen­tic­ate them­selves like users do. (neces­sary if user inform­a­tion is not rep­lic­ated).

If user authen­tic­a­tion inform­a­tion is rep­lic­ated, they should be stored at sites where the users access the sys­tem.

Dis­trib­uted author­iz­a­tion rules are stored in the cata­logue. Hand­ling user groups for authen­tic­a­tion sim­pli­fies dis­trib­uted data­base admin­is­tra­tion.

Full rep­lic­a­tion of author­iz­a­tion has 2 advant­ages:
 — Sim­pler
 — Can be done at com­pile time

How­ever dis­ad­vant­age is that there is a higher over­head cost.

Dis­trib­uted Semantic Integ­rity Con­trol

The main prob­lems in design­ing an integ­rity sub­sys­tem for a dis­trib­uted DBMS are:
 — Defin­i­tion and stor­age of asser­tions
 — Enforce­ment of these asser­tions

Defin­i­tion of Dis­trib­uted Integ­rity Asser­tions

3 classes of asser­tions:
 — Indi­vidu­al asser­tions: single-rela­tion single-vari­able asser­tions.
 — Set-ori­ented asser­tions: single-rela­tion mul­tivari­able con­straints
 — Asser­tions involving aggreg­ates: requires spe­cial pro­cessing due to cost of eval­u­at­ing the aggreg­ates.
New integ­rity asser­tion: Sites that store the rela­tions involved in the asser­tion.

An integ­rity asser­tion becomes a dis­trib­uted oper­a­tion in 2 steps:
 — Trans­form high-level asser­tions into com­piled asser­tions
 — Store com­piled asser­tions accord­ing to class of asser­tion.

Enforce­ment of Dis­trib­uted Integ­rity Asser­tions

Main prob­lem is to decide at which site to enfor­ce the integ­rity asser­tions.

Choice depends on:
 — Class of asser­tion
 — Type of update
 — Nature of site where update is issued.


“ —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  — – — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — – — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — –

Appoġġ traduzz­joni: http://amzn.to/1Z7d5oc
 — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — – — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — – — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — –

Kon­troll tad-Dejta Semantika

kon­troll tad-data semantiku tipika­ment tinkludi l-ġest­joni fehma, kon­troll tas-sig­urtà u l-kon­troll integ­rità semantika. Dawn il-fun­z­jon­iji­et jiżgur­aw li l-utenti awtorizza­ti jwettqu operazz­jon­iji­et kor­retta fuq il-bażi tad-dejta. Dan jikkon­trib­wixxi għall-integ­rità data­base.

Fehmi­et fil DBMS Imqas­sam

Veduta f’database mqass­ma ma tkunx ġej­ja mid-relazz­jon­iji­et fram­ment­ati maħżuna f’siti dif­fer­enti. Meta ħsieb huwa defin­it, l-isem tagħha u mis­toqsi­ja irkupru tagħha huma maħżuna fil-kata­logu.

L-immap­pjar ta mis­toqsi­ja espress fuq opin­jon­iji­et isir bl-istess mod bħal fis-sis­temi ċen­t­ralizza­ti. Fil sis­temi ċen­t­ralizza­ti dan isir per­mezz ta modi­fika mis­toqsi­ja. A mis­toqsi­ja mqass­ma hija pproċes­sata mill-proċes­sur mis­toqsi­ja mqass­ma.

Fehmi­et fi DBMS dis­trib­witi jista jsir bl-użu regoli simili għar-regoli ta’ definizz­joni fram­ment. Il-proċes­sur mis­toqsi­ja mapep tal-mis­toqsi­ja mqass­ma fi query fuq fram­menti fiżiċi.

Opin­jon­iji­et miksuba mid- data­base mqass­ma tista tkun għal­ja biex jeval­wa. View derivazz­joni isir bil­li jingħaq­du il-kwal­i­fika fehma mal-kwal­i­fika mis­toqsi­ja. L-altern­at­tiva hija li tuża Snap­shots. Daqqa t’għajn lejn jir­rappreżenta stat partikolari ta data­base. Dan ifis­ser li huwa statiku. Snap­shots huma utli meta l-utenti ma jkunux interessa­ti li jaraw l-aktar ver­ż­joni reċenti tad-data­base. Aċċess per­mezz Snap­shots huwa ade­g­wat għall-mis­toqsiji­et li huma stabbil­iti minn qabel.

Kon­troll Awtorizza­zz­joni Imqas­sam

Aktar kump­lessi minn f’databases ċen­t­rali minħab­ba l-fatt li l-oġġet­ti u l-utenti jist­għu jitqass­mu. Dawn il-prob­lemi huma:
 — Awtentikazz­joni utent Remote
 — Ġest­joni tar-regoli ta awtorizza­zz­joni mqass­ma
 — Żam­ma ta fehmi­et u grup­pi ta’ utenti.

Biex jip­pre­vjenu l-aċċess utent remot mill-utenti mhux awtorizza­ti, 2 approċċi huma użati:
 — Inform­azz­joni ta utenti awtorizza­ti huma maħżuna fuq is-siti koll­ha fil-katal­gu (iktar għaljin)
 — Is-siti koll­ha ta ‘DBMS dis­trib­witi jaw­tentikaw lil­hom infushom bħall-utenti do. (Meħtieġ jekk l-inform­azz­joni l-utent ma jkunx rep­likati).

Jekk l-inform­azz­joni awtentikazz­joni utent huwa rep­likat, dawn għand­hom jiġu maħżuna f’siti fejn l-utenti jaċċes­saw is-sis­tema.

regoli ta awtorizza­zz­joni dis­trib­witi huma maħżuna fil-katal­gu. Hand­ling grup­pi ta utenti għall-awtentikazz­joni jis­sim­pli­fika l-ammin­is­trazz­joni dejtabejż mqass­ma.

rep­likazz­joni sħiħa tal-awtorizza­zz­joni għan­du 2 vantaġġi:
 — sem­pliċi
 — Jista jsir fi żmi­en jikkom­pilaw

Madankollu iżvantaġġ huwa li hemm spiża over­head ogħla.

Imqas­sam Kon­troll Integ­rità semantika

Il-prob­lemi ewlen­in fit-tfassil ta sub­sistema integ­rità għal DBMS dis­trib­witi huma:
 — Definizz­joni u ħażna ta affer­mazz­jon­iji­et
 — Infurz­ar ta dawn l-affer­mazz­jon­iji­et

Definizz­joni-affer­mazz­jon­iji­et Integ­rità Imqas­sam

3 klassiji­et ta affer­mazz­jon­iji­et:
 — Affer­mazz­jon­iji­et individwali:-relazzjoni unika affer­mazz­jon­iji­et wieħed-var­jab­bli.
 — Orjentata lejn is-Set affer­mazz­jon­iji­et: restrizz­jon­iji­et-relazz­joni unika mul­tivari­able
 — Dikjarazz­jon­iji­et li jin­volvu l-aggreg­ati: jir­rikjedi speċ­jali ta ipproċes­sar minħab­ba l-ispiża ta’ eval­wazz­joni tal-aggreg­ati.
Ġod­da affer­mazz­joni integ­rità: Siti li jaħżnu r-relazz­jon­iji­et invol­uti fl-affer­mazz­joni.

Affer­mazz­joni integ­rità ssir operazz­joni dis­trib­witi fi 2 passi:
 — Trans­form allegazz­jon­iji­et ta liv­ell għoli fis affer­mazz­jon­iji­et kkumpil­ata
 — Aħżen ikkumpil­ata allegazz­jon­iji­et skond il-klassi ta affer­mazz­joni.

Infurz­ar tal Dikjarazz­jon­iji­et Integ­rità Imqas­sam

prob­lema prinċip­ali hija li tid­deċiedi fuq liema sit biex tin­forza l-affer­mazz­jon­iji­et integ­rità.

Għażla tid­de­pendi fuq:
 — Klassi ta affer­mazz­joni
 — Tip ta Aġġor­na­ment
 — In-nat­ura tas-sit fejn aġġor­na­ment jinħareġ.


“ —  —  —  —  —  —  —  —  —  —  —  —  —  —  —  — – — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — – — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — –

Onder­steun­ing ver­tal­ing: http://amzn.to/1Z7d5oc
 — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — – — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — – — —  —  —  —  —  —  —  —  —  —  —  —  —  —  — –

Semantiese Data Con­trol

Semantiese data beheer sluit tipies oog bestuur, sekur­iteit beheer en semantiese integ­riteit beheer. Hierdie funk­sies verseker dat gemagtig­de gebruikers uit te voer kor­rek bedry­wighede op die data­bas­is. Dit dra by tot die data­bas­is integ­riteit.

Uit­sig in Dis­trib­uted DBMS

N Blik in n ver­spreide data­bas­is is afgelei van gefrag­men­teer­de ver­houd­ings ges­toor op ver­skil­lende plekke. Wan­neer die oog word gedefinieer, sy naam en sy her­win­ning nav­raag word ges­toor in n kata­logus.